We found that adult prostate stem cells (APSCs) reside in the proximal region of ducts, are quiescent, produce high levels of active TGF-? and are the targets of malignant transformation. By comparing gene expression profiles of APSCs and fetal prostate stem cells (FPSCs) to that of mature cells we found that APSCs display a signature of quiescence, while FPSCs a signature of rapid proliferation. Of 6 genes we found to be uniquely over expressed in APSCs, two transcription factors, KLF4 and JUNB, and an enzyme, peptidylarginine deiminase type IV (PADI4), all with previously suggested context-dependent tumor suppressor or oncogene function, were examined for their role in normal stem cells and in prostate cancer. Interrogation of tumor- mRNA from a cohort of 129 prostate cancer patients revealed that 25 (19%) over-express KLF4 and JUNB and that this group of patients remained relapse-free for up to 10 years post-surgery. PADI4, which we show controls APSC homeostasis by increasing active TGF-?, had no prognostic value in this cohort. However, knockdown (KD) of each of these 3 genes in APSCs increased stem cell proliferation and altered their differentiation while KD in AKT-transformed cells increased tumorigenicity. KD of KLF4 and JUNB induced epithelial-mesenchymal transition (EMT). Since KLF4 and JUNB predict recurrence-free prognosis in prostate cancer patients and since all 3 genes have profound effects on stem cell homeostasis and have anti-tumor effects in the mouse prostate, we hypothesize that they are an integral part of a network that impedes prostate cancer progression. We propose to test this hypothesis in two aims. In Aim 1 we will examine the mechanism by which PADI4 and KLF4 (individually and together), control cell cycle and stem cell homeostasis. Using mass spectroscopy, we will determine whether PADI4 increases TGF-? activity by blocking methylation of crucial arginines on an inhibitory Smad, Smad7, and by increasing expression of ?vb6 integrin, an activator of TGF-?, creating a positive feedback loop of TGF-? activation. We will also generate a mouse with a prostate epithelium-specific conditional KLF4 knockout to determine directly the effect of KLF4 on stem cell fate and prostate morphology in resting and cycling prostate. These experiments will elucidate the mechanisms by which PADI4 and KLF4 regulate APSC homeostasis. In Aim 2 we will examine whether the anti-tumor effect of PADI4, KLF4 and JUNB in human and mouse tumors with active AKT is mediated through their regulation of tumor initiating cells (TICs) in culture and in vivo. We will also determine whether KLF4 and JUNB proteins, like their mRNA, are biomarkers of good prognosis. This will be done in the original cohort of 129 patients and validated by extending this analysis to 544 additional prostate cancers. This proposed broad approach to study the role of PADI4, KLF4 and JUNB in the homeostasis of APSCs in vitro and in vivo, their anti-tumor role in the function of TICs and the utility of KLF4 and JUNB protein detection for prognosis of human prostate cancer, will result in significant advances to our understanding of the mechanisms that go awry during oncogenesis.